Detecting collusion in procurement auctions

TL;DR

This paper develops a machine learning model to detect collusion in procurement auctions with 91% accuracy, analyzing Russian auction data and interpreting decision factors, contributing to anti-collusion efforts.

Contribution

A novel machine learning approach with interpretability for detecting auction collusion, validated on real auction data and behavior analysis.

Findings

Model predicts collusion with 91% accuracy.

Interpretability via Shepley vector elucidates decision factors.

Honest company behavior confirmed through simulation.

Abstract

The study aimed at detecting cartel collusion involved analyzing decisions of the Russian Federal Antimonopoly Service and data on auctions. As a result, a machine learning model was developed that predicts with 91% accuracy the signs of collusion between bidders based on their history after dividing 40 auctions into test and training samples in a 30/70 ratio. Decomposition of the model using the Shepley vector allowed the interpretation of the decision-making process. The behavior of honest companies in auctions was also studied, confirmed by independent simulation validation.